Economists Do It With Models

Warning: “graphic” content…

Bookmark and Share
Confirmation Bias (And Stereotypes) In Action…

May 26th, 2009 · 9 Comments
Behavioral Econ · Discrimination · Gender

Sometimes the posts write themselves, really. From a Facebook reader:

“I was totally surprised to see a girl pic when I got to this page! I came across your writing today for the first time and after a few articles I imagined a dude writing them… I guess that just shows how some people (maybe the majority) grow up with certain biases, in this case a gender bias towards males, (more biases: racist, religious, geographic, etc.) You could use this as a topic if you you haven’t yet… worth mentioning I thought.”

I initially planned to just reply to this individual via my Facebook page, but as I was writing I decided that this reader was right and his issue did in fact deserve a post. (Or maybe I was just overreacting to the fact that I come across as a dude on the interwebs, since that has much potential for creepy.)

I assume that we are all familiar with the concept of a stereotype, more technically referred to in the social sciences as statistical discrimination. Stereotyping occurs when we impart the typical characteristics of a group on an individual because that individual is a member of that group. Let me give a topical example: In general, girls like Barbies. As my reader above unexpectedly learned, I am a girl. Stereotypes would suggest that I like Barbies. I use this example to illustrate the point that this line of reasoning doesn’t logically hold, since I for one have never quite understood the entertainment value of a skinny plastic doll that doesn’t talk, light up, dance, set itself on fire, or really do much of anything. (In fairness, I think some of them are awesomely pretty and am happy to let them sit around in their boxes. I feel like there is a metaphor to be had there. Add in the fact that I was all about My Little Pony when I was a kid and you have much fodder for a potential therapist, who knows.) Lesson learned: just because a group shows a characteristic in general does not mean that any particular member of that group exhibits that characteristic.

Stereotyping explains why my reader initially assumed I was a guy- the reality of the world is that most economists are male, and he (this reader is male) certainly had a higher probability of being right by guessing that I was a guy. In this way, he made a good guess, and as long as he keeps in mind that it’s only a guess, we aren’t dealing with a severe bias in judgment. But let’s continue…

There’s this interesting psychological concept called confirmation bias whereby people tend to seek out and interpret information in a way that supports their existing hypotheses and ignore information that doesn’t. Given that, here is how my reader’s thought process likely went:

  • I think this author is a guy.
  • Notice: I see that this person writes about sports.
  • Ignore: A male would probably not use a magenta color as part of his web site color scheme.
  • Notice: There is a decent amount of sarcasm/snark going on in the writing, which is more representative of a male writer.
  • Ignore: The author’s screen name for comments is econgirl. (Come on now.)
  • Notice: Would a girl really use a play on words about doing it with models?
  • Ignore: I have a post, currently on the front page, entitled “Women In Science: I Am Doing My Part, Apparently…”
  • And so on…

People do this all the time without even realizing it. Consider the following task, which is the example often given in classrooms to illustrate the concept of confirmation bias: you are given the numbers 2, 4, 6 and told that you need to figure out what the rule for generating those numbers is. The way you figure out the rule is by suggesting other sequences of 3 numbers and having a judge tell you yes or no. The vast majority of people will ask about sequences that are likely to give them an answer of yes rather than sequences that yield an answer of no, even though the no answers are usually more helpful.

In the above example, the underlying rule is “sequences of increasing positive numbers”. People usually start with a hypothesis of “consecutive even numbers” or something like that. As such, the game often goes as follows:

  • 8, 10, 12: yes
  • 20, 22, 24: yes

And so on. What would give a lot more insight would be to challenge the hypothesis of consecutive even numbers, since you could rule it out as a possibility pretty easily:

  • 3, 6, 9: yes
  • 1, 5, 90: yes

See how you learned a lot more from the second set of tests than the first?

Why is this relevant to economics? Economics is at least partly about understanding what (and how) people choose to purchase and consume, and these decisions are based on what people know about the products that are available to them. If people are biased in how they process information about what to consume, they are also going to be biased in their consumption decisions, and that fact is certainly economically relevant. It’s also relevant from a marketing perspective, though in that case the lesson is to try to take advantage of the confirmation bias and remember just how important a first impression can be!

For the record, this is the picture that the reader came upon on Facebook:

It pleases me because I’m sort of giving a look that says “Really? You thought I was a guy?…hmph.”

Tags: Behavioral Econ · Discrimination · Gender

9 responses so far ↓

  • 1 Mele // May 26, 2009 at 12:17 pm

    Good post. Another common test of confirmation bias in psychology is to show subjects a set of four cards, showing “A”, “B”, “2”, and “3”. (Wason, 1966) Next, tell them that all vowels have an even number on the back. Then, ask them to test the theory by picking two cards to turn over. Most people will select “A” and “2”. Though “A” is a good choice because it can both confirm and deny, “3” is the scientifically logical second choice.

  • 2 Lance // May 26, 2009 at 12:25 pm

    I thought that look was “Really? You thought I was older?…hmph.”

  • 3 econgirl // May 26, 2009 at 8:46 pm

    @ Mele: That is another of my favorite tests.

    @ Lance: Ha. I have a Dorian Gray-esque portrait of myself under my bed getting older as we speak. =P

  • 4 Pablo // May 27, 2009 at 6:51 am

    Good Stuff…
    You are totally right on the thought process. I can’t believe I even ignored “econgirl” and just kept seeing the facts that would confirm my bias. You know what I like about this blog the most? The brainstorming that it can lead to… Ever since I left college I feel more and more disconnected with books, professors, knowledge, etc. and reading your posts make feel closer to that. My brainstorming got me to another interesting topic derived from the gender issue, affirmative action. Whether you are in favor or against this (I am definitely against even though I have and could further benefit from it) there are economics behind it. It would definitely be a negative impact in terms of numbers (loss of GDP, loss of Total surplus,etc), but it may be a gain as a whole? (social inclusion). I’d like to see economics being a little more focused on human beings and their well being rather than just human beings and their “rational” choices regarding consumption…

    Glad to have been a source of inspiration …

  • 5 Unormal // May 27, 2009 at 12:13 pm

    lol, good try, but everyone knows there are no girls on the internet.

  • 6 econgirl // May 27, 2009 at 1:09 pm

    “On the Internet, nobody knows you’re a dog.”

    http://www.unc.edu/depts/jomc/academics/dri/idog.html

  • 7 Krzysztof Wiszniewski // Jun 12, 2009 at 7:55 pm

    This is why “About The Author” is the first thing I read on any new blog (that and finding things to be snide about if I don’t like what they’re writing… 😉 )

  • 8 Eben A Knowlton // Jul 26, 2012 at 9:39 am

    Visited Sears yesterday and saw an industrial sized tool box..all colored PINK! Let the thinking/stereotypes begin! (My internal reaction was, “WHY?”

  • 9 Jeff // Aug 9, 2013 at 11:24 am

    Mele – What about a card with “B” on one side and “E” on the other?

Leave a Comment